-
i
Abstract:
Imprinting is the process by which precocial species learn the
characteristics of their parents, and
associative learning occurs when animals link an unconditioned
stimulus (UCS) to a conditioned stimulus
(CS). Yet what is the relationship between imprinting and
associative learning? This has been
discussed using two major assumptions in psychological
literature. Firstly imprinting runs parallel to
associative learning. The two processes are distinguished on
basis of the specificity of imprinting, purpose,
and brain localisation. Most investigations into imprinting
based on this assumption give conclusions which
support a complete dissociation of imprinting and associative
learning, excepting the PAL phenomenon
which may link the localisation of the two processes.
Secondly, investigations which assume imprinting can be
interwoven with associative learning on different
levels have discussed their relationship in terms of imprinting
being solely a form of associative or
recognition learning, yet concrete conclusions cannot been made
due the appearance of recognition
learning elements within an associative learning model. Thus,
attempts to singularly label imprinting as
either linked or dissociated with associative learning (in
assuming the two are parallel to one another), or
that imprinting is either a form of associative or recognition
learning (in assuming the two are interwoven)
have not been successful.
Another model which similarly assumes the interweaving of
imprinting and associative learning has
surfaced, yet unlike previous theories, does not seek to
characterise imprinting in terms of a singular banner,
and thus does not facilitate comparison with existing learning
processes. It is able to link imprinting
theories based on the two major assumptions by accounting for
phenomena observed in their investigations.
The only problem is that this model cannot generate predictions
of imprinting. Thus, until this can occur,
we must rely on and seek to further research into theories which
characterise imprinting solely in terms of
one theory.
299 words
-
ii
Table of Contents:
I. Abstract i
II. Table of Contents ii
III. Introduction 1
a. What is filial imprinting? 1
b. What is associative learning? 1
IV. Imprinting and associative learning running in parallel
2
a. Distinction in specificity of imprinting 2
b. Distinction in purpose 2
c. Distinction in localisation of function 3
V. Imprinting as interwoven with associative learning 4
a. Imprinting as a form of associative learning 4
b. Searching for a neurological unconditioned stimulus 6
c. Imprinting as a combination of associative learning and
recognition learning 7
VI. Conclusion 10
VII. References 11
-
1
Introduction:
What is filial imprinting?
As humans, an essential part of our development is learning to
recognise and attach to our parents or
guardians. This phenomenon is also seen, and much more
significant in the development of non-human
animals. Zebra young are able to distinguish the stripes of
their parents from a herd of 100; zoo-keepers
must consider the time a mother and new-born baby spend apart so
the baby panda is able to attach to the
correct mother figure; or, if their parents are absent, geese
are able to imprint on human companions, and
cases of geese learning how to fly by direction from ultralight
aircraft powered by enthusiast Bill Lishman
have been recorded in the movie Fly Away Home. This attachment
to a parental carer is called imprinting,
and is the phenomenon by which an animal learns to recognise
their parental and conspecific (of same
species) characteristics, and has both filial (relationship with
parent) and sexual (in choice of later mate)
consequences.
This essay will address the learning process underpinning filial
imprinting, which was initially observed as
the phenomenon in which precocial birds (species of birds whose
young are relatively mature or mobile
from moment of hatching) elicit a following response toward the
first object they see whether this be a
member of their own species or a foreign object, e.g. a human or
a rotating red box. This effect can be
replicated in an artificial setting through what is called a
discrimination task, where a chick is presented
with two visual stimuli (this can be simultaneously or in
succession) one is an object with which it has
been previously exposed and trained with, another is a novel
object. Imprinting has successfully occurred
to the trained object if the chick elicits specific behaviours
such as the following response (the chick
simply follows the object), or emits twitters toward this
object. Objects such as flashing lights are often
used as an imprinting object, as substitute for a stuffed fowl,
as they elicit the same responses in chicks.
What is associative learning?
Associative learning (a form of which is classical conditioning)
can be defined as the linkage of an
unconditioned stimulus (UCS) to a conditioned stimulus (CS),
such that the CS comes to predict the UCS.
This carries a strong element of reinforcement, whereby the UCS
is either negative or positive for the
learner, and the CS is associated with a good or bad experience.
On a very simple level, associative
learning can be tested for by rewarding a chick for choosing a
specific visual pattern (CS), and testing if
the preference exists in the absence of reinforcement (UCS).
Differentiating between learning processes is important in
discovering its boundaries, thus furthering
knowledge about how imprinting occurs and how a relationship
between a parent and baby forms, for
Comment [TC1]: Introduction: What is filial imprinting?
Comment [CT2]:
http://en.wikipedia.org/wiki/File:Christian_Moullec_4.jpg
Comment [CT3]: Use quote from Lorenz original paper?
Comment [CT4]: Add to explanation with introduction from
no.9
Comment [TC5]: What is associative learning?
http://en.wikipedia.org/wiki/File:Christian_Moullec_4.jpghttp://en.wikipedia.org/wiki/File:Christian_Moullec_4.jpg
-
2
potential benefit in a zoological context. There are many papers
which discuss the interplay between
imprinting and associative learning, but a direct comparison of
all these theories has not been achieved,
thus, the research question is, what is the relationship between
imprinting and associative learning?
Imprinting and associative learning running in parallel:
Distinction in specificity of imprinting
Konrad Lorenz (1937) first discovered imprinting, and argued for
its uniqueness, meaning it had no relation
to associative learning apart from being categorised under the
broad umbrella of learning. In this way, he
directly compared the two processes, assuming they occurred in
parallel. He proposed 4 aspects in which
imprinting was more specific than associative learning there was
a critical period in which it occurred, it
was irreversible, an outcome of imprinting was not observed
until later in life, and the object imprinted
upon was not that to which the ultimate outcome of imprinting
was made. However, at that time, sexual
and filial imprinting had not been separated, yet now in the
analysis of filial imprinting alone, the last two
features described are not applicable, as the perception of the
outcome of imprinting as being the choice
of mate later in life stem from the idea that imprinting is a
continuous process from filial recognition to
sexual selection. This has now been adjusted to restrict the
outcome of filial imprinting to be the filial
response detailed above, and later sexual selection is
excluded.
The argument of a critical period has also been refuted in
Sluckin and Salzen (1961), which demonstrated
that the ending of an animals critical period can be prolonged
by restricting the chicks early exposure to
stimulation, and the argument of irreversibility has been
disproven in studies (Salzen & Meyer, 1968)
showing that an imprinted preference can be written over. Thus,
the argument that imprinting can be
differentiated from associative learning due to its specificity
has little merit. As points of similarity between
the two processes are not identified when merely considering the
specificity of imprinting, for the purpose
of discussing the relationship between imprinting and
associative learning, Lorenzs argument will not be
considered.
Distinction in purpose
On a surface level, imprinting is also compared in terms of its
purpose. It is argued that the purpose of
imprinting is to learn conspecific characteristics, which,
according to Bateson (1990) is an example of
learning to categorise; however associative learning enables
predictions about the environment ..the
job of learning to predict and control the environment is not
the same as that of learning to categorize it.
Comment [TC6]: Idea of basic?
Comment [TC7]: Imprinting and associative learning running in
parallel? Distinction in specificity of imprinting
Comment [TC8]: Lorenz, K 1937, The Companion in the Birds World,
Journal fr Ornithologie, Vol. 54, p.264-5
Comment [TC9]: Possibly in no.8?
Comment [TC10]: refereneced
Comment [CT11]: However this is a very simplistic
distinction
Comment [TC12]: No.5.. referenced
Comment [CT13]: Patrick Bateson (6 P89) Detecting causual
structure may required classification, but establishing a
classification does not involve an association of cause with
effect
Comment [TC14]: After this, there was A point of difference is
also identified in the significance of the order of presentation of
stimulus when learning to categorise information, order is
insignificant, however when learning cause and effect, the cause
must precede the effect, not vice versa. -- but thats somewhat
futile because in imprinting, cause (CS) and effect (UCS) both
appear at same time.. OH, but thats the argument - but its too
early to appear here havent discussed associative learning yet
Comment [CT15]: Same argument apparently implicit in
.Associative learning is like a procedure, recognition memory is
more declarative implicit on P124 Memory, Imprinting, and the
Brain
-
3
This argument also takes the same line as an evolutionary
analysis, whereby two memory systems are
disassociated based on difference in function. This theory is
proposed by Sherry and Schacter (1987),
arguing that the systems controlling different learning
processes are dissociated when they are specialized
to such a degree that the functional problem each system handles
cannot be handled by another system. In
practice, if imprinting and associative learning can handle the
problems of the other, then the occurrence
of imprinting should improve associative learning and
associative learning should improve imprinting,
however this is not the case in Bolhuis and Johnson (1988) and
Bateson and Reese (1968). Thus,
imprinting and associative learning can be dissociated in
comparison of purpose and function.
Distinction in localisation of function
Neural studies comparing the two processes have also argued for
their dissociation. Such investigations rest
on the assumption that if these two processes are similar, they
will be localised to the same areas. Initially,
an area of the brain called the Intermediate part of the Medial
Hyperstriatum Ventral (IMHV) was
discovered to be involved in imprinting through the accumulation
of metabolic markers such as radioactive
uracil (Horn et al., 1979), and more recently experimental
studies lesioning the chick IMHV have also
confirmed the localisation of imprinting to this area (McCabe et
al. 1981 and 1982, and Takamatsu &
Tsukada, 1985).
For example, in a study conducted by McCabe et al. (1982),
bilateral (left and right hemisphere) lesions
were placed in the chick IMHV 3 hours after exposure to an
imprinting stimulus. There were 3 other
experimental groups in the study some chicks had hyperstriatum
accessorium (HA) or lateral cerebral
area (LCA) lesions, and there was also a sham-operated control
group (undergoes the same experience as
an operation, but no lesions are placed). 15-20 hours after the
operation, they were tested in a
discrimination task and all the chicks except for those with
lesions in the IMHV showed a preference for
the imprinting stimulus.
One of the criticisms of using lesions to test for a difference
between an experimental and control group is
the presence of a confounding variable in the potential
destruction of motor or sensory capability as caused
by the lesion. Therefore, if a chick had indeed imprinted or was
able to imprint, yet could not elicit the
response necessary to show their preference for the imprinting
stimulus, they would have failed the
discrimination task. However, after the operation, all the
chicks used in this study performed equally well
in tests of visuomotor co-ordination, ruling out such an
explanation of the findings.
Upon discovering where imprinting is localised, its relationship
with associative learning based on a shared
localisation has been disproven in various studies, including
the aforementioned McCabe et al. (1982). The
chicks with IMHV lesions were tested in a task which required
associative learning they were able to
distinguish between two visual patterns when rewarded for making
the correct choice. Associative
learning had occurred in the absence of imprinting, proving
their disparity.
Comment [TC16]: In no.10 in references
Comment [TC17]: Horn, G. & Johnson, M.H. 1989, Memory
Systems in the Chick: Dissociations and Neuronal Analysis,
Neuropsychologia, Vol. 27 (1), p.6
Comment [TC18]: Distinction in localisation of function
Comment [TC19]: Weiskrantz no.10 says its not just localisation
p.8
Comment [CT20]: CHECK
Comment [TC21]: In Horn. G. 1991 .(in references)
Comment [CT22]: ((An interesting characteristic of the role of
the IMHV in imprinting is the accumulation of experimental evidence
suggesting the differential roles of the right and left IMHV
Cipolla-Neto et al. (1982) (or another clearer study with just
unilateral or bilateral) placed either right or left lesions in
chicks approximately 3 hours after exposure to an imprinting
stimulus. 15-20 hours after the operation, all the chicks (with
either right or left lesions) retained their preference for the
imprinting stimulus, however after a second contralateral (in the
other hemisphere) lesion was placed in the chicks, only those which
had the first lesion placed in the right hemisphere passed the
discrimination task.))
Comment [TC23]: In Horn, G. 1991, Neural bases of recognition
memory, Behavioural and Neural Aspects of Learning and Memory,
Oxford Science Publications, p.
Comment [TC24]: Where else?? - references
Comment [TC25]: Species of chick?
Comment [TC26]: Maybe just say lesions in other parts of
brain
Comment [CT27]: WHERE IS THE STUDY SAYING IF THIS WAS A READ-OUT
MECHANISM OR OTHER??
Comment [CT28]: BUT Johnson and Horn (1986) did not consider
this to be conclusive evidence (11)
1.Simultaneous discrimination may be easier than successive
discrimination tasks i.e. method in testing imprinting and
associative learning was different for associative learning, both
objects were presented simultaneously but in imprinting they were
presented one after other (although because all groups went through
this the same problem should have affected all but I dont know if
the difference between groups was significant or not) ...
-
4
This conclusion (that imprinting and associative learning are
fundamentally different) then diminishes in
credibility through the appearance of a process considered a
form of associative learning passive
avoidance learning (PAL). PAL is process of acquiring an
avoidance reaction to an object due to negative
reinforcement. This process was tested by taking advantage of a
chicks spontaneous tendency to peck at a
bright bead and thus allowing them to peck at a bead coated in
metylanthranylate (MeA), which has an
aversive taste. Patterson et al. (1990) and Davies et al. (1988)
used different methods to inhibit the
functioning of the chicks IMHV; however both tested whether
these chicks were able to associate the
aversive taste experienced with the bead. It was found that the
chicks without a functioning IMHV did not
refrain from pecking the bead, yet sham-operated controls or
checks with other areas of the brain inhibited
did. Thus, it appeared there was a form of associative learning
that was localised to the same area of the
brain as imprinting.
However, there have been some explanations of why this result
occurred, centred on the difference between
the task being tested and associative learning. Firstly, the
methodology of the experiments are questioned,
as they only involved one-trial of the task and adequate
exposure to the unpleasant taste of the bead was not
achieved - chicks may have refrained from pecking after a few
more trials.
Secondly, it can be argued that PAL is different to associative
learning. Inhibition from pecking a bright
bead requires declarative recognition due to it being the
reverse of the chicks spontaneous tendency to
peck. All other associative learning tasks tested natural
responses such as looking for food or recognising
visual patterns, however the credibility of these explanations
has not been tested.
Thus, theories assuming imprinting as a process running parallel
to associative learning has offered mainly
evidence against a close relationship between these two
processes, however there may be some possibility
for a link in terms of the form of associative learning, PAL,
being localised to the same region of the brain
governing imprinting.
Imprinting as interwoven with associative learning:
The other main view of imprinting is that it can be considered
in terms of other learning frameworks, and
not directly compared to them the two main frameworks are
imprinting as a form of recognition learning
or associative learning. In some literature, perceptual learning
(van Kampen, 1993) or exposure learning
(Sluckin, 1973) has been used instead of the term recognition
learning, however both have the essential
characteristic of arguing imprinting is the learning of an
animal to recognise their parents, involving the
formation of a mental representation without reinforcement. For
the purpose of this essay, recognition
learning will be as a blanket term for such theories.
The distinction between these two frameworks is
well-characterised in the following quote from van
Comment [CT29]: Also called acquisition of a negative
preference
Comment [TC30]: In G.Horn no.12?
Comment [CT31]: (before the training of the PAL task)
Comment [TC32]: Cite study
Comment [TC33]: Another word?
Comment [TC34]: Imprinting as a layer relationship?? Imprinting
as recognition or associative learning A distinction in terms of
purpose
Comment [TC35]: True, yes?
-
5
Kampen (1993) The perceptual-learning hypothesis of filial
imprinting assumes that the level of arousal
[when perceiving the imprinting object] affects the formation of
an internal representation of the imprinting
object, while the associative-learning hypothesis assumes that
the internal representation also becomes
associated with the physiological state of the moment, that is,
the reinforcing properties of the
unconditioned stimulus function by altering the chicks
physiological state.
The assumption underlying the analysis of recognition learning
arguments is that, if imprinting is
interpreted under such a framework, then the relationship
between imprinting and associative learning is
limited to a mutual categorisation under the broad umbrella of
learning. It should be noted that some
theorists do not agree with this, van Kampen (1993) argues that
associative learning theory does not
necessarily contradict recognition learning theory, they can be
interwoven.
Imprinting as a form of associative learning
There are several arguments for the consideration of imprinting
in terms of an associative learning or
recognition learning framework. The most researched argument is
that imprinting can be understood
through an associative learning framework, and imprinting is a
type of associative learning in which a
chick associates a conditioned stimulus (CS) with an
unconditioned stimulus (UCS). Hoffman and his
colleagues (1972, 1973a, 1973b, 1978, 1983, 1987) developed a
theory of what the conditioned and
unconditioned stimulus in imprinting would be the conditioned
stimulus is called E1, (the conspicuous
characteristics of the imprinting object), and the E2 is an
unconditioned stimulus, possibly a
motivationally significant event experienced as a direct result
of imprinting.
The major problem with this interpretation is that upon
receiving stimulus of the imprinting object, a chick
would experience both E1 and E2, and the object acts as its own
reinforcing stimulus. This makes the
proposed model of imprinting difficult to investigate, although
scientists have tested if phenomena
exhibited as a result of associative learning can also be
exhibited in imprinting this would mean
imprinting carries an essential associative learning principle.
These phenomena include overshadowing and
blocking. Overshadowing occurs when several E1s (CS) are paired
with one E2 (UCS); the E1s which have
been presented separately from the others is more strongly
associated with the E2 than E1s which have
been presented as part of a compound stimulus (many E1s at once)
the E1s presented separately are said
to overshadow the others. Blocking occurs when an E1 that has
already been paired with an E2 before
being presented as part of a compound stimulus blocks the
association of the other members of the
compound stimulus from being associated with the E2.
Experiments employing this paradigm have had some success. For
example, de Vos and Bolhuis (1990)
tested for blocking by conducted an experiment with two
conditions one group of chicks was exposed to
two yellow cylinders (YY), another group was exposed to two red
cylinders (RR) the prior exposure.
Comment [CT36]: Create a link to behavioural explanation here,
to make clear Perhaps reinforce that perceptual learning theory
doesnt refute associative learning theory? the second is an
extension of the first? linked below
Comment [CT37]: van Kampen p.118
Comment [CT38]: Should I list them, saying four? YES, I think
only 2 though?
Comment [TC39]: Do principles of associative learning also
underlie imprinting?
Comment [CT40]: Give all the years P 317 (4) **** ask about
referencing
Comment [CT41]: the most successful of which is ______
Comment [TC42]: Join together somehow if needed
Comment [CT43]: Also say others have been theorized (movement,
however it was disproven as still can imprint on stationary objects
(P 317-318 (4), low anxiety state (operant conditioning Moltz, 1960
P324 (4) has this gotten anywhere)
Comment [TC44]: Note from plan: The imprinting object is the
reinforcing stimulus in itself (creates problems when trying to
test it) Patrick Bateson (__) , and also Campbell, B.A. and
Pickleman, J.R. (1961) Hoffman?
Comment [CT45]: Are the explanations clear enough?
Comment [TC46]: referenced
-
6
After six days of exposure, the cylinders for both groups were
replaced for seven days by one blue and one
yellow cylinder (YB). After all this, when both groups were
exposed to a single blue cylinder (B), the RR
group spent more time with it. This suggests that when exposed
to the YB cylinders, the YY group had
blocked an association with the blue cylinder as they had
already be exposed to a yellow one. In a similar
manner, overshadowing was reported in a study conducted by
Hoffman et al. (1972).
In evaluation, these experiments have a major flaw the apparent
result of blocking or overshadowing may
merely be due to incidentally more attention given to one or the
other stimulus when experiencing a
compound stimulus the stimuli were physically disparate and
chicks could only concentrate on one at a
time. Solutions to this have been suggested (Bolhuis, de Vos
& Kruijt, 1990) in the form of presenting a
combination of auditory and visual stimuli have been suggested,
so that the chick does not need to give
selective attention to one or the other as a chick does not stop
listening to one stimulus in order to listen to
another. However, such experiments employing this methodology
have not been conducted.
Searching for a neurological unconditioned stimulus
Yet, these experiments, simply establish the existence of a UCS,
but what not what it is, which is where
neural interpretations become important. Hoffman and Segal
(1983) theorised the presence of a pleasure
centre co-ordinating the motivationally significant event. Thus,
the unconditioned stimulus is the
stimulation of the pleasure centre. This theory can be supported
if filial response can be heightened by
increasing the intensity of the imprinting stimulus. This has
been shown in a study by Jackson and Bateson
(1974) whereby changing the colour of a visual stimulus
presented to chicks that were predisposed to react
to certain colour, changes the likelihood they would step onto
pedal to be exposed to the imprinting
stimulus (the filial response). However, this depended on the
length of pre-exposure to the coloured
stimulus.
It is, however, not known what chemicals would stimulate this
pleasure centre. It was suggested that the
catecholamine neurons worked as reward neurons. More
specifically, Davies, Horn and McCabe (1983)
found that exposure to artificial imprinting stimulus was
connected with widespread elevation of
noradrenaline level in the chick brain, but not with variations
in dopamine level. Furthermore, depleted
levels of noradrenaline in the chick forebrain increase the time
taken for imprinting to occur (Davies,
Johnson & Horn, 1992).
However, by increasing when imprinting occurs, the role of
neurotransmitters as reinforcers support both
an associative learning and recognition learning framework for
imprinting. White and Milner (1992)
theorised that reinforcers had the potential to 1. To enhance
the formation of neural representations. Here,
the purpose of the reinforcer is to give arousal, enabling
memory formation, thus being closer to a
recognition learning interpretation of imprinting. 2. To endow a
stimulus with motivational properties,
Comment [CT47]: Significantly??? Have a look** (7)
Comment [TC48]: referenced
Comment [CT49]: Although mixed in with some unexpected ones P322
(4)
Comment [CT50]: What does this mean for the results?? Bolhuis,
De Vos and Kruijt suggested combination of auditory and visual
stimuli - have there been any studies??? Look** Called the
sensory-orientation hypothesis p.113 van Kampen
Comment [TC51]: referenced
Comment [CT52]: ******Also use transfer of training/stimulus
generalisation as evidence (Bateson P96 (6))??
Comment [TC53]: Searching for a more detailed explanation a
neurological UCS?
Comment [TC54]: referenced
Comment [TC55]: referenced
Comment [CT56]: Studies found in vanKampen p.117 referenced
Comment [CT57]: Also supported by this study: Dudai (1989)
Noradrenaline has been shown to facilitate long0term potentation
(LTP), a neuronal phenomenon suggested to be the first stage in
mammalian learning. It has been shown that NMDA receptors and
glutamate are involved in LTP in the mammalian hippocampus, and
also in neural changes in the IMHV related to imprinting. These
parallel findings lead to the suggestion that the first stages of
memory formation in imprinting may well involve LTP-like cellular
changes, and, one step further, that they may be modulated by
noradrenaline. van Kampen p.119
Comment [TC58]: referenced
-
7
which would facilitate the formation of an association between
the imprinting stimulus and the appropriate
response, closer to an associative learning interpretation of
imprinting. Yet, despite the potential presence
of recognition learning, neurotransmitters are still considered
reinforcers in a CS-UCS context reinforcers
which would enhance recognition learning. Recognition and
associative learning would appear on different
levels in imprinting.
Thus, the position of each neurotransmitter in terms of
behavioural models is ambiguous. Research has
been conducted as to the effects of neurochemical factors on
imprinting as a whole, yet research into the
pathway by which this effect is triggered is yet to surface; and
perhaps the pathway is not solely
recognition or associative but a combination of both on various
levels.
Imprinting as a combination of associative learning and
recognition learning
This sort of an argument, combining aspects of both recognition
(perceptual) and associative learning, so
associative learning would be a component of imprinting is
proposed in the Bateson and Horn (1994)
model of imprinting. Essentially, it proposes three
sub-processes that sequentially analyse the features of
stimuli, recognize those feature combinations that are familiar
and organize appropriate responses. Those
three sub-processes are Analysis, Recognition and Executive
respectively. Figure 1 shows a simple
version of the model where there are 2 Analysis modules, 2
Recognition modules and 1 Executive module.
The overall model is a neural net, and there would be many more
modules forming a layer of Analysis
modules connected to a layer of Recognition modules, etc. in a
real chick.
Figure 1:
Caption: A module is represented by a single square. A system is
composed of many modules. from Bateson and
Horn (1994)
Analysis modules act as detectors of the features of a stimulus
and work to decompose the sensory input
into a variety of features, such as lines, orientation,
direction of movement and colour. One Analysis
module corresponds to one feature, and also transmits onwards
the intensity with which each of the
Comment [TC59]: Can also add A further problem in investigating
these pathways is that the distinction between them is becoming
increasingly blurred some researchers have even suggested all
learning Is associative learning. Scientific consensus must be
reached before any further research occurs in this area.
Comment [TC60]: Imprinting as a combination of recognition and
associative learning.
Comment [TC61]: referenced
Comment [CT62]: * But it is considered to be a perceptual
learning theory P122 van Kampen But also incorporates elements of
perceptual learning which are [expand]
Comment [CT63]: Mention thickness of lines.. (give a key)
Comment [TC64]: Caption: A module is represented by a single
square. A system is composed of many modules.
-
8
feature detectors has been activated, or in other words how
strongly a detected feature corresponds with
the chicks store of the imprinted stimulus.
This transmits onto a Recognition module. A Recognition module
consists of an excitatory unit (P-unit),
and an inhibitory (I-unit) but this discussion will only feature
excitatory units in discussing the formation of
associations. Bateson and Horn (1994) say A key feature of our
model is the spontaneous [i.e. random]
fluctuation of excitability in the plastic (P) units of the
Recognition system and the excitability of the P
units within different Recognition modules are not synchronised.
This is based on the finding of
spontaneous neural activity in the IMHV of both imprinted and
non-imprinted chicks as detected by a
micro-electrode in Payne & Horn (1984), Davey & Horn
(1991) and Brown & Horn (1994). When a P unit
is at a high level of excitability, a given input from an
Analysis module is most likely to activate it, and this
connection between a unit in the Analysis module and the P unit
in the Recognition module is strengthened
when the two units are conjointly active; the more frequent this
conjoint activation occurs, the stronger the
connection is herein lies the fundamental associative learning
principle. Although the random excitability
of the P units of the Recognition module makes it seem unlikely
a strong association can form to a specific
P unit, it is proposed that a P unit remains in a heightened
excitability for a finite period after activation,
allowing for such a connection to form.
Modules in the Executive system control the filial approach
behaviour of chicks and can be activated by
modules in the Analysis system or Recognition system. Units in
the Executive module initially (before
imprinting) hold a maximum connection with units in the Analysis
module, representing the predisposition
of chicks to approach certain objects rather than others.
However, as a result the conjoint activation of a P
unit in the Recognition module and a unit in the Analysis
module, resulting in the activation of an
Executive module, the 3 connections between them are all
strengthened, and the Recognition modules
rapidly gain effective control of the Executive module, enabling
control over the filial approach
behaviour.
Comment [TC65]: referenced
Comment [TC66]: referenced
Comment [TC67]: Possibly cite source which also says this
Comment [TC68]: Word better?
Comment [CT69]: accounting for?? similar word..
Comment [CT70]: Still need to cover the inactivation rule and
the Recognition modules storing information and the likelihood of
Analysis modules to activate Recognition modules (talk about
together)
-
9
Figure 2:
Caption: Diagram (a) in this figure shows what happens after
exposure to a stimulus that activated Analysis module,
A1. The spontaneous excitability of the P unit in recognition
module, R2, happened to be higher than that in R1 at
the time when the input from A1 arrived, so the units in A1 and
R2 were conjointly active, increasing the strength of
the connection between them (represented by the thickened
lines). At the same time, the state of excitation of a P unit
in the Executive module E was enhanced by the direct link with
A1, and as a result of conjoint activation with the P
unit in R2, the connection between them is also
strengthened.
Diagram (b) in this figure illustrates what happens after the
process of imprinting is complete, and the A1-R2 and R2-
E connections are at full strength and the A2-R2 and A2-E
connections have disappeared, however have the potential
to re-appear if they are conjointly active imprinting is
reversible. from Bateson and Horn (1994)
In evaluation, this model has been deemed supportive of
associative learning by van Kampen and de Vos
(1995) as the premise that the connection between the
recognition and the executive system is
strengthened as a result of conjoint activityis what is
generally accepted to the basis of all associative
learning. However in van Kampen (1993) it was heralded as a
perceptual-learning model, (i.e.
recognition learning) as it carried elements of recognising and
storing representations of the imprinting
object. Thus, the model includes certain aspects of both
recognition and associative learning concepts. This
model is able to account for phenomena observed in a solely
associative-learning interpretation of
imprinting (e.g. blocking, overshadowing), and takes PAL in an
imprinting context. Furthermore, evidence
used to support dissociation between imprinting and associative
learning when assuming they run parallel
to one another can also be accommodated in this theory. The fact
that associative learning occurs and
imprinting cannot when in the IMHV is lesioned may be due to the
IMHV accommodating the
Recognition modules, however the Executive modules would still
be activated and join with the
Analysis modules. A major problem, however, is the little
experimental evidence to justify the model and
that the learning rules are far too obscure to allow for
specific predictions to be derived from them.
Comment [TC71]: van Kampen, H.S., de Vos, G.J. 1995, A Study of
Blocking and Overshadowing in Filial Imprinting, The Quarterly
Journal of Experimental Psychology, Vol. 48B (4), p.353
Comment [TC72]: ***why??
Comment [TC73]: Say by performing classification-together effect
or too complicated? Idea reference: Bateson what must be known in
order to understand imprinting
Comment [CT74]: But can the evidence for this (i.e.
overshadowing and blocking) still apply to this model?
A model of the classification-together effect has been
constructed by Patrick Bateson account for what happens in the real
world benefit of greater accuracy no. 6 p.89-90
But are they the same thing?? Not really classification-together
effect enables phenomenon of overshadowing but it is the
overshadowing that is more important in showing that UCS exists.
Although, the very fact that a classification-together (more like
classification-apart) effect can be applied to imprinting means
that a chick can imprint to 2 conditioned stimuli at one time, thus
enabling it to be tested out. is this important in terms of a
`development of a model of how imprinting occurs though??
Comment [TC75]: Possibly reference the section ^?
Comment [TC76]: Or system
Comment [TC77]: Say that interpretation that imhv lesions show
NOT associative learning is not true (from no.12 p.225)
Comment [TC78]: vanKampen p.122
-
10
Conclusion:
There have been two major types of theories modelling imprinting
assuming imprinting is parallel and
can be directly compared to associative learning, or imprinting
as being interwoven with associative
learning on various levels. The Bateson and Horn (1994) model is
a stand-out from the second category as
it explains both an associative-learning-alone interpretation of
imprinting (also accounting for PAL), and
findings of lesion studies assuming a parallel between
imprinting and associative learning, thus linking
these two types of theories. In this way, purely for the
purposes of characterising imprinting, this model can
be considered superior.
Yet in terms of contextualising imprinting, when directly
comparing imprinting and associative learning
and assuming the two processes run in parallel, most evidence
points to a distinction between the two;
when assuming imprinting is interwoven with associative
learning, most research displays a foundation on
the basic associative learning principle, however is unclear if
this entails that imprinting is a form of
associative learning, or if there are additional elements (i.e.
recognition learning) perhaps overshadowing
the salience of associative learning as the primary learning
process underpinning imprinting, or if
imprinting should incorporate both associative learning and
these other elements.
This problem is primarily due to the models theorised being too
complex for application of a simple
definition processes of recognition and associative learning can
be observed in various levels within these
models. Therefore, it is perhaps not wise to attempt to
categorise imprinting under a single banner.
Although this would prevent a simple contextualisation of
imprinting in terms of existing learning theories,
it is not all-important in developing an understanding of
imprinting itself. What is more important is an
accurate model which is able to account for phenomena studied
and generate predictions, which is partially
provided by Bateson and Horn (1994). Thus, unless the Bateson
and Horn (1994) model can provide
reliable predictions of imprinting, the best method of
understanding the phenomenon is by comparing it to
associative and recognition learning processes which, through
comparisons to humans, we have greater
knowledge.
Comment [CT79]: This is essentially my analysis can I use I?
Comment [TC80]: Apart from PAL
Comment [TC81]: So imprinting can be characterised as
recognition learning
-
11
References:
Primary Sources
Bateson, P.P.G. 1990, Is imprinting special?, Philosophical
Transactions of the Royal Society of London,
Vol. 329, No. 1253, p.129.
Bateson, P.P.G. & Horn, G. 1994, Imprinting and recognition
memory: a neural net model, Animal
Behaviour, Vol. 48, pp695-700.
Bolhuis, J.J., de Vos, G.J. & Kruijt, J.P. 1990, Filial
Imprinting and Associative Learning, The Quarterly
Journal of Experimental Psychology, Vol. 42B (3), p.324.
de Vos, G.J. & Bolhuis, J.J. 1990, An investigation into
Blocking of Filial Imprinting in the Chick During
Exposure to a Compound Stimulus, The Quarterly Journal of
Experimental Psychology, Vol.42B (3),
p.289.
Lorenz, K 1937, The Companion in the Birds World, Journal fr
Ornithologie, Vol. 54, pp.264-265.
McCabe, B.J., Cipolla-Neto, J., Horn, G. & Bateson, P.P.G.
1982, Amnesic effects of bilateral lesions in
the hyperstriatum ventral of the chick after imprinting,
Experimental Brain Research, Vol. 48, pp.13-21.
Salzen, E.A. & Meyer, C.C., Reversibility of Imprinting,
Journal of Comparative and Physiological
Psychology, Vol.66 (2), pp.269-275.
Sluckin, W. 1973, Imprinting and Early Learning, 2nd Edition,
Aldine Transaction, New Jersey, pps.108-
111.
van Kampen, H.S. 1993, An analysis of the learning process
underlying filial imprinting, University of
Groningen, Groningen, pp111-127.
van Kampen, H.S. & de Vos, G.J. 1995, A Study of Blocking
and Overshadowing in Filial Imprinting,
The Quarterly Journal of Experimental Psychology, Vol. 48B (4),
p.353.
Secondary Sources
Bateston, P.P.G & Reese, E.P. 1968, Reinforcing properties
of conspicuous objects before imprinting has
occurred, Psychonomic Science, Vol. 10, pp379-380. In Horn, G.
& Johnson, M.H. 1989, Memory
Systems in the Chick: Dissociations and Neuronal Analysis,
Neuropsychologia, Vol. 27 (1), p.6.
-
12
Bolhuis, J.J. & Johnson, M.H. 1988, Effects of
response-contingency and stimulus presentation schedule
on imprinting in the chick, Journal of Comparative Psychology,
Vol. 102 (1), pp.61-65. In Horn, G. &
Johnson, M.H. 1989, Memory Systems in the Chick: Dissociations
and Neuronal Analysis,
Neuropsychologia, Vol. 27 (1), p.6.
Brown, M.W. & Horn, G. 1994, Learning-related alterations in
the visual responsiveness of neurons in a
meory system of the chick brain, European Journal of
Neuroscience, Vol. 6 (9), pp1479-1490. In Bateson,
P.P.G. & Horn, G. 1994, Imprinting and recognition memory: a
neural net model, Animal Behaviour, Vol.
48, p699.
Davey, J.E. & Horn, G. 1991, The development of hemispheric
asymmetries in neuronal activity in the
domestic chick after visual experience, Behavioural Brain
Research, Vol.45, pp81-86. In Bateson, P.P.G.
& Horn, G. 1994, Imprinting and recognition memory: a neural
net model, Animal Behaviour, Vol. 48,
p699.
Davies, D.C., Horn, G. & McCabe, B.J. 1983, Changes in
telencephalic cathecolamine levels in the
domestic chick. Effects of age and visual experience, Brain
Research, Vol. 312 (2), pp.251-255. In van
Kampen, H.S. 1993, An analysis of the learning process
underlying filial imprinting, University of
Groningen, Groningen, p117.
Davies, D.C., Horn, G. & McCabe, B.J. 1992, The effect of
the neurotoxin DSP4 on the development of a
predisposition in the domestic chick, Developmental
Psychobiology, Vol.25 (4), pp251-259. In van
Kampen, H.S. 1993, An analysis of the learning process
underlying filial imprinting, University of
Groningen, Groningen, p117.
Hoffman, H.S. 1978, Experimental analysis of imprinting and its
behavioral effects. In Bower, G. (editor),
The Psychology of Learning and Motivation, Vol. 12, pp1-37. In
Bolhuis, J.J., de Vos, G.J. & Kruijt, J.P.
1990, Filial Imprinting and Associative Learning, The Quarterly
Journal of Experimental Psychology,
Vol. 42B (3), p.317.
Hoffman, H.S. 1987, Imprinting and the critical period for
social attachments: Some laboratory
investigations. In Bornstein, M.H. (editor), Sensitive periods
in development, pp.99-121. In Bolhuis, J.J.,
de Vos, G.J. & Kruijt, J.P. 1990, Filial Imprinting and
Associative Learning, The Quarterly Journal of
Experimental Psychology, Vol. 42B (3), p.317.
Hoffman, H.S. & Ratner, A.M. 1973a, Effects of stimulus and
environmental familiarity on visual
imprinting in newly-hatched ducklings, Journal of Comparative
and Physiological Psychology, Vol. 85,
pp11-19. In Bolhuis, J.J., de Vos, G.J. & Kruijt, J.P. 1990,
Filial Imprinting and Associative Learning,
The Quarterly Journal of Experimental Psychology, Vol. 42B (3),
p.317.
-
13
Hoffman, H.S. & Ratner, A.M. 1973b, A reinforcement model of
imprinting: Implications for socialization
in monkeys and men, Psychological Review, Vol. 80, pp.527-544.
In Bolhuis, J.J., de Vos, G.J. & Kruijt,
J.P. 1990, Filial Imprinting and Associative Learning, The
Quarterly Journal of Experimental Psychology,
Vol. 42B (3), p.317.
Hoffman, H.S., Ratner, A.M. & Eiserer, L.A. 1972, Role of
visual imprinting in the emergence of specific
filial attachments in ducklings. Journal of Comparative and
Physiological Psychology, Vol. 81, pp.399-409.
In Bolhuis, J.J., de Vos, G.J. & Kruijt, J.P. 1990, Filial
Imprinting and Associative Learning, The
Quarterly Journal of Experimental Psychology, Vol. 42B (3),
pp.317, 322
Hoffman H.S. & Segal, M. 1983, Biological factors in social
attachments: A new view of a basic
phenomenon. In Zeiler, M.D. & Harzem, P. (editors), Advances
in analysis of behaviour, Vol. 3, pp-41-61.
In Bolhuis, J.J., de Vos, G.J. & Kruijt, J.P. 1990, Filial
Imprinting and Associative Learning, The
Quarterly Journal of Experimental Psychology, Vol. 42B (3),
p.317.
Hoffman H.S. & Segal, M. 1983, Biological factors in social
attachments: A new view of a basic
phenomenon. In van Kampen, H.S. 1993, An analysis of the
learning process underlying filial imprinting,
University of Groningen, Groningen, p114.
Horn, G., McCabe, B.J. & Bateson, P.P.G. 1979, An
autoradiographic study of the chick brain after
imprinting, Brain Research, Vol.168, pp361-373. In Horn, G.
1990, Neural bases of recognition memory,
Philosophical Transactions of the Royal Society of London,
Series B, p.135.
Jackson, P.S. & Bateson, P.P.G. 1974, Imprinting and
exploration of slight novelty in chicks, Nature,
Vol.251, pp609-610. In van Kampen, H.S. 1993, An analysis of the
learning process underlying filial
imprinting, University of Groningen, Groningen, p115.
McCabe, B.J. Horn, G. & Bateson, P.P.G. 1981, Effects of
restricted lesions of the chick forebrain on the
acquisition of filial preferences during imprinting, Brain
Research, Vol. 205, pp.29-37. In Horn, G. 1990,
Neural bases of recognition memory, Philosophical Transactions
of the Royal Society of London, Series
B, p.135.
Payne, J.K. & Horn, G. 1984, Long-term consequences of
exposure to an imprinting stimulus on
spontaneous impulse activity in the chick brain, Behavioural
Brain Research, Vol. 13, pp155-162. In
Bateson, P.P.G. & Horn, G. 1994, Imprinting and recognition
memory: a neural net model, Animal
Behaviour, Vol. 48, p699.
-
14
Sherry, D.F. & Schacter, D.L. 1987, The evolution of
multiply memory systems, Psychological Review,
Vol. 94, pp439-454. In Weiskrantz, L. 1990, Problems of learning
and memory: one or multiple memory
systems?, Philosophical Transactions of the Royal Society of
London, Series B, p.105.
Sluckin, W. & Salzen, E.A. 1961, Imprinting and perceptual
learning, Quarterly Journal of experimental
Psychology, Vol.55, pp181-187. In Sluckin, W. 1973, Imprinting
and Early Learning, 2nd
Edition, Aldine
Transaction, New Jersey, p.82.
Takamatsu, K. & Tsukada, Y. 1985, Neurobiological basis of
imprinting in chick and duckling. In
Tsukada, Y. (editor), Perspectives on neuroscience from molecule
to mind, pp.187-106, Berlin: Springer-
Verlag. In Horn, G. 1990, Neural bases of recognition memory,
Philosophical Transactions of the Royal
Society of London, Series B, p.135.
White, N.M. & Milner, P.M. 1992, The psychobiology of
reinforcers, Annual Review of Psychology,
Vol.43, pp443-471. In van Kampen, H.S. 1993, An analysis of the
learning process underlying filial
imprinting, University of Groningen, Groningen, p119.